{"data":{"slug":"deadbits-vigil-llm","name":"vigil-llm","tagline":"Detect prompt injections and other risky inputs in LLMs","github_url":"https://github.com/deadbits/vigil-llm","owner":"deadbits","repo":"vigil-llm","owner_avatar_url":"https://avatars.githubusercontent.com/u/1332757?v=4","primary_language":"Python","stars":496,"forks":56,"topics":["adversarial-attacks","adversarial-machine-learning","large-language-models","llm-security","llmops","prompt-injection","security-tools","yara-scanner"],"archived":false,"github_pushed_at":"2024-01-31T18:43:41+00:00","maintenance_label":"Dormant","stars_delta_30d":5,"url":"https://www.graphcanon.com/tools/deadbits-vigil-llm","markdown_url":"https://www.graphcanon.com/tools/deadbits-vigil-llm.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/deadbits-vigil-llm","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=deadbits-vigil-llm","description":"⚡ Vigil ⚡  Detect prompt injections, jailbreaks, and other potentially risky Large Language Model (LLM) inputs","homepage_url":"https://vigil.deadbits.ai/","license":"Apache-2.0","open_issues":16,"watchers":9,"ai_summary":"Vigil is a tool designed for detecting prompt injections and jailbreak attempts aimed at Large Language Models, falling under the category of Evaluation & Observability.","readme_excerpt":"## Install Vigil 🛠️\n\nFollow the steps below to install Vigil\n\nA [Docker container](docs/docker.md) is also available, but this is not currently recommended.\n\n---\n\n### Install YARA\nFollow the instructions on the [YARA Getting Started documentation](https://yara.readthedocs.io/en/stable/gettingstarted.html) to download and install [YARA v4.3.2](https://github.com/VirusTotal/yara/releases).\n\n---\n\n### Install Vigil library\nInside your virutal environment, install the application:\n```\npip install -e .\n```","github_created_at":"2023-09-04T17:02:21+00:00","created_at":"2026-07-07T17:43:34.135088+00:00","updated_at":"2026-08-21T06:01:46.576977+00:00","categories":[{"slug":"evaluation-observability","name":"Evaluation & Observability","url":"https://www.graphcanon.com/categories/evaluation-observability","markdown_url":"https://www.graphcanon.com/categories/evaluation-observability.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/evaluation-observability"}],"tags":[{"slug":"adversarial-attacks","name":"adversarial-attacks"},{"slug":"large-language-models","name":"large language models"},{"slug":"llm-security","name":"llm security"},{"slug":"prompt-injection","name":"prompt-injection"}],"trust":{"provenance":{"is_fork":false,"github_id":687122549,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-08-21T06:01:45.611Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":932,"last_release_at":"2023-12-31T17:18:52Z","stars_delta_30d":5,"open_issues_delta_30d":0},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T11:23:56.085Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-08-21T06:01:46.248Z"},"deploy":{"source":"dockerfile:Dockerfile","self_host":true,"observed_at":"2026-08-21T06:01:46.248Z","managed_saas":false},"languages":{"value":["python"],"source":"github.language","observed_at":"2026-08-21T06:01:46.248Z"},"has_docker":{"value":true,"source":"dockerfile:Dockerfile","observed_at":"2026-08-21T06:01:46.248Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-08-21T06:01:46.248Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When deploying large language models that require high levels of input security, vigil-llm can be employed to detect maliciously crafted inputs intended to manipulate model behavior.","In environments where adversarial machine learning is a risk, vigil-llm's capability to identify prompt injections makes it an essential tool for safeguarding LLMs."],"when_not_to_use":["If your application does not require high security against malicious inputs or if the risks of prompt injection are minimal due to controlled input sources, vigil-llm might be unnecessary.","For projects that focus on optimizing output speed rather than input robustness, other tools might be more appropriate as vigil-llm could add significant processing overhead."],"source":"enrich:decision_facts","observed_at":"2026-07-15T09:47:59.738Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"Vigil-llm is designed for users who need robust security measures to protect against prompt injections and jailbreak attempts in LLMs."}]}}